Perinatal antiretroviral (ARV) medications lead to dramatic reductions in mother to child transmission (MTCT) of HIV-1, but in utero exposure to combination ARVs may adversely affect neurodevelopment. Children born to HIV-infected women may also be at generally higher risk for neurodevelopmental delay and mortality than children born to HIV-uninfected women in resource limited settings. At least 10 million HIV-uninfected children have been born to HIV-infected mothers over the past 10 years, and the use of ARVs in pregnancy is being rapidly scaled up globally. It is important to determine the consequences of both perinatal HIV-1 and ARV exposure, in order to help inform the rational selection of specific ARVs for MTCT prevention or allay existing concerns about their use, and to improve overall health outcomes among HIV-exposed children. In Botswana, one in three pregnant women is infected with HIV-1, reflecting a generalized epidemic that affects all social strata. The Botswana Harvard School of Public Health AIDS Initiative Partnership (BHP) has conducted clinical research related to MTCT prevention and maternal/pediatric health in Botswana since 1999. We are uniquely positioned to simultaneously study the impact of perinatal ARV- and HIV exposure on child health and neurodevelopment. Our primary objectives are: 1. To compare neurodevelopmental outcomes following in utero exposure to 3-drug HAART that contains 2 or more nucleoside reverse transcriptase inhibitors (NRTIs), vs. monotherapy with a single NRTI (zidovudine), among HIV-uninfected children born to HIV-infected mothers with similar antenatal CD4+ cell counts. 2. To prospectively compare neurodevelopmental outcomes among HIV-exposed but uninfected children, vs. children born to HIV-uninfected women. 3. To prospectively compare the rates of mortality in HIV-uninfected children born to HIV-infected mothers vs. children born to HIV-uninfected women in a setting of HAART availability. 4. To prospectively determine the biologic and social factors that are most strongly associated with excess mortality and neurodevelopmental delay among HIV-exposed but uninfected children compared with HIV-unexposed children. To address these aims, we will prospectively enroll, follow, and test (until 2 years) two new cohorts: a cohort of 400 HIV-uninfected children born to HIV-infected mothers, as well as 400 children born to HIV-uninfected mothers, in Botswana. In addition, to address Aim 1, we will also perform neurodevelopmental testing (at 2 years of age) on a subset of a third, existing cohort: 730 children who were exposed perinatally to HAART in an ongoing clinical trial at these same locations.